Nonlinear Relationships between Birds and Vegetation

Abstract

We tested for the presence and importance of nonlinear relationships among bird species abundances and vegetation characteristics of the lower Colorado River valley. We transformed vegetation variables into polynomials and then ran multiple regressions, with bird densities as the dependent variables. Polynomial variables were selected in 36% of first steps and 45% of second steps in the regressions. In a significant number of cases there was a response to a polynomial variable but not to the original linear variable. Importance of curvilinear responses varied among bird species. This method offers a more comprehensive examination of responses to envirnomental gradients because unimodal and some skewed relationships can be detected.